New vapor pressure equations from triple point to critical point and a

New vapor pressure equations from triple point to critical point and a ... for the Vapor Pressure of Pure Components Constructed on Characteristic Poi...
0 downloads 0 Views 1MB Size
J. Chem. Eng. Data 1906, 31 438-447

430

New Vapor Pressure Equations from Triple Point to Critical Point and a Predictive Procedure for Vapor Pressure G. R. SomayaJulu Thermodynamics Research Center, Texas A& M University, College Station, Texas 77843

The quadratic and the cubic equatlons, previously proposed by us, for the representation of vapor pressure from the triple point to the normal boiling point, have now been extended to the critical polnt by the addltlon of two more parameters to each of them. They have been compared with the equations proposed by Goodwln and Wagner. The extended quadratlc equatlon employing only four parameters has been found to represent the vapor pressure of even such a complex substance as water qulte adequately. Finally, a new predictive procedure has been proposed to predict the vapor pressure from normal boiling to crltical polnt even when no vapor pressure data are present.

The new equation (4 parameters): In (P/P,) = AX

Numerous procedures have been proposed for the representation of vapor pressure from the triple point to the critical point. Some of them have been recently reviewed by Wagner ( 7 ) and by Ambrose (2). While various organizations are still compiling Antoine constants, and use different Antoine equations for different regions of the liquid substances, the ESDU (3) uses one simple equation containing not more than 5 parameters to represent the vapor pressure for the entire liquid region. Experience has shown that such a procedure is not only expedient but is also the most accurate one. This paper is aimed at improving such procedures. We begin this by defining certain symbols we encounter in our presentation. T , = triple point T , = normal boiling point T, = critical point T , = inversion point, the point where the deviations from a linear equation derived on the basis of the vapor pressure data at the normal boiling point and critical point changes sign as shown in Figure 1. I n case the deviations do not change sign, but show an inflection, T , will be referred to as the inflection point. r , = r , or r , T , = T I T , (reduced temperature) P , = triple point pressure f b = 1 atm (101.325 kPa) P , = critical pressure P , = P, or Pb P , = pressure at the inversion point x = ( r , - r ) / T ;X , = ( r ,- r ) / r ;xb= ( r , - r ) / r Y = ( r c- r ) / r ;z = ( r , - r)/r,; w = r,/r G = ( r , - T,)/r,; H = ( r , - r,)/r, u = standard deviation - Pm,&)/PeXpt,= relative deviation (P ur = standard deviation based on relative deviations We now introduce the following equations The quadratic equation (3 parameters): In (P/P,) = AX

+ BX2

(1)

The cubic equation (4 parameters): In (P/P,) = AX

+ BX2 + C X 3

(2)

0021-9568/86/1731-0438$01.50/0

(3)

The new equation (5 parameters): In ( P / P ), = AX

+ BX2 + CX3 + X(DZ' +€P) (4)

The new equation (4 parameters): In ( P / P , ) = AY

+ BY2 + C,? + D P

(5)

The new equation (5 parameters):

In (P/P,) = AY

+ B Y 2 + CY3 + D Z ' +

€2

(6)

The Wagner equation (4 parameters): In (P/P,) = AY

Introduction

+ BX2 + X(CZ' + D P )

+ W(Bz' + C P + Dzh)

(7)

The Wagner equation (5 parameters): In ( P / P , ) = A Y

+ W(Bz'+

CP

+ Dzh+

€2) (8)

The Goodwin equation (4 parameters): In ( P / P ), = A(X/G)

+ B(X/G)' + C ( X / G ) 3+ D ( X / G ) ( Y / H ) ' (9)

The Goodwin equation (5 parameters): In (P/P,) = A ( X / G )

+ B ( X / G ) ' + C ( X / G ) 3+ D(X/G)4

+ € ( X / G ) ( Y / H ) ' (IO)

The A , B , C , D , and € of eq 1- 10 are the coefficients to be determined; f , g, h , and i are the exponents either predetermined or to be determined through nonlinear regressions or through iteration. The eq 1 and 2 were introduced by us previously (4) for representing vapor pressure from the triple point to the normal boiling point. T, is the hidden parameter in these two equations. T , is supplied and is not considered as an additional parameter in the eq 3, 4, 9, and 10. I n 1 and 2, we may have any boiling point T , between the triple point and the normal boiling point, and it may be determined through regression, either linear or nonlinear, as shown by us previously. The equations will be referred to as, for example, la, or 1b, depending upon whether the T, is T , or Tb. By adding two addiiional parameters to each of the eq 1 and 2, we obtain the eq 3 and 4. The same extensions will be used for the eq 2, 3, and 4. The eq 3 and 4 reduce to zero at T = T, but at T = T , they reduce to the eq 11 and 12, respectively:

+ BG2 = AG + 6 G 2 + CG3

In (P,/P,) = AG In (P,/P,)

(11) (12)

The new eq 3 and 4 are related to the Goodwin equations (9) and (IO) and to the Wagner equation (7). Our investigation reveals that at least two Gocdwin terms and two Wagner terms are needed to make the vapor pressure equation a reliable one. The new equations, eq 5 and 6, are also formulated on the basis of the quadratic and cubic equations but are constrained to go through the critical point. They are closely related to the Wagner equation (7). The extensions a and b also apply to the Goodwin equations (9) and (10). While constrained to produce

0 1986 American

Chemical Society

Journal of Chemical and Engineering Data, Vol. 31, No. 4, 1986 430

-50t

-100

-1501 0.5

I

I

0.6

0.7

I

0.8

I

0.9

I

1.0

REDUCED TEMPERATURE, Tr

Flgure 1. Plot of the pressure deviations vs. reduced temperature for water, ethanol, and argon. The pressure deviations are the dlfferences

between the observed pressure and the pressure calculated from the equation In P = a b l T . The coefficients a and b were evaluated by using the pressure and temperature data points at the normal bolllng point and the critical point.

-

the required pressures at T, the eq 5, 6,7,and 8 are constrained to produce the selected critical pressure. I t is much more important to produce better results in the low-pressure region than in the critical region. I n this respect eq 3,4,9, and 10 are superior to the eq 5, 6,7,and 8. The quadratic equation is most useful for the prediction of vapor pressure from the triple point to the normal boiling point as shown by us previously. The cubic equation plays an important role in the determination of the normal boiling point. Equations 3, 5 , 7,and 9 contain Only 4 parameters and appear to be quite adequate for the representation of vapor pressure of most compounds. The 5-parameter analogues of these equations namely, (4).(6),(8), and (lo),are necessary only when much more accuracy in the representation of vapor pressure data is needed. All the equations (3)-(10)satisfy the following criteria: 1. d2PldT2= at T,. 2. A H l A Z is minimum at T, = 0.85-0.90. 3. A H l A Z = (1lj3) (1 T,)" below the normal boilingpoint. d2PldT2should be indeterminate at the critical point (5)-(8). The criterion 2 was first pointed out by Waring ( 9 ) ,who noticed that AHlAZfor water passes through a minimum at T, = 0.85. AH is the enthalpy of vaporization and AZ is the change in the compression factor on evaporation; AZ = 2, - Z, = (V, VJPIRT, V, and V, being the molar volumes of vapor and liquid, respectively. A H l A Z is related to the vapor pressure by the equation

-

AH/AZ = RT2 d In P/dT

(13)

which is derived from the Clapeyron equation. The third criterion follows from the Watson relation (IO). The value of K usually lies between 0.33 and 0.4as shown by Gambill ( 11) but the value of 0.375as recommended by Ambrose, Counsell, and Hicks (12) appears to be quite satisfctory. j3 is usually a constant for a given substance in the region between the triple point and the normal boiling point. I n addition to the above criteria, we also have three additional criteria with respect to a function a, introduced by Plank and Riedel (73)and defined as

a = (dP/dT)T/P

(14)

Although Plank and Riedel were concerned with the value of

a at the critical temperature, we find its value to be useful at all temperatures from the triple point to the critical point. The behavior of a is particularly interesting in the critical region, since it goes to a minimum at Tr = 0.97-0.98.This behavior of the a-function may be seen from the plots of a vs. temperature made by Wagner ( 1) for both argon and nitrogen in the region close to the critical point. We have shown in Figure

t 5

1

90

' 1 ' 1 100 110

' 1 ' 1 ' 1 ' 1 120 130 140 150

' I ' " ' ' 1 ' I

1

160 170 180 190 2 0 0

TEMPERATURE, K

Figure 2. Plot of the a-function for methane vs. temperature from the

triple point to the critical polnt.

2 a plot of the a-function for methane from the triple point to the critical point. This gives rise to the following criterion. 4. a = minimum at T, = 0.97-0.98. We have not found so far any exception to this new criterion. At the critical temperature, a becomes equal to -A of the eq 5 , 6,7,and 8. The mathematical expressions for cy derived on the basis of eq 1-10 are given in Table X I I I . The expressions for the a based on eq 3-10 lead to the following criterion. 5 . daldT = a at T., From eq 13 and 14 and the criterion 3,one can write for a the following expression

= (l/PRT)(l - T,)"

(15)

which satisfies the criterion 5, providing the value of K is positive and less than unity. We may also express a alternatively as follows: a=a+bX

+ b X + cX2 a = exp(a + bT + cT2) a = exp(a + bT, + CT;) cy

=a

a =a

+ bZ'+

c P + dzh

(16) (17) (18) (19) (20)

The eq 16-19 do not satisfy the criterion 5 but may be used to express a as a function of temperature below the critical temperature. Equation 18 actually forms the basis of the Cox equation discussed by us previously (4). Equation 20 would satisfy the criterion 5 provided the value of f is positive and less than 1. I n principle, we should be able to formulate a good vapor pressure equation on the basis of the a-function. A sixth criterion may also be established with respect to the a-function. This has to do with the value of a at T = T,. I t can be easily verified that the constant A of the eq 1 and 2 is equal to -a at T = T,. The most satisfactory value of a is that obtainable from the cubic equation (2) between the triple point and the normal boiling point. The sixth possible criterion, therefore, is that the value of a obtainable from any of the eq 3-10 at T, be in agreement with the value obtainable from the

440 Journal of Chemical and Engineering Data, Vol. 3 1, No. 4, 7986 I

I

I

I

I

I

I

I

I

I

I

I

z 0 c 5 > W

0 W

2

c

4W U

A

n

A

0

-0.001

0

t " Figure 3. Plot of the relative deviations obtained on the basis of the 4-parameter eq 3a, 3b, 5, 7, 9a, and 9b for water below the normal boiling temperature.

cubic equation or for that matter from any other equation which can represent the vapor pressure data in the low-pressure region most accurately. I f this criterion were not to be satisfied then either the vapor pressure equation or the vapor pressure is suspect. The tests conducted by us have indicated that the 4-parameter equations are, in general, defective in the lowpressure region. The disagreement may also be due to the vapor pressure data in the low-pressure region. For cases for which the triple point pressure is unavailable, it is better to estimate the triple point pressure as suggested by Ambrose and Davies (74) and use it in the regression in order to obtain a satisfactory value for a at T,. The 5-parameter equations are most reliable at the triple point.

Results and Discusdon According to Scott and Osborn (75), the reliability of a vapor pressure equation depends upon its adequacy in the representation of the vapor pressure of such a complex substance as water. For this reason, we have selected water as the typical substance for testing the adequacy of the selected vapor pressure equations in this article. Using the vapor pressure data of water taken from the most reliable sources (76-22), we have found that all the equations, (3), (5), and (7), containing only 4 parameters are quite satisfactory. For the purpose of this paper, all the points are given equal weight except the critical point which is excluded. The critical point was taken from the NBSINRS Steam Tables (22).The triple point pressure was taken from the work of Johnson, Guiklner, and Jones (27)to be 0.61173 kPa at 273.16 K. The c r i t i i l pressure was determined by us through iteration to be 22060 kPa. This value is slightly different from the value 22055 kPa recommended by the NBSINRS Steam Tables (22).The 5-parameter eq 4, 6, 8, and 10 are possibly the best for the representation of the vapor pressure data for water. Although the Goodwin eq 9 is satisfactory for most substances, it is not adequate for water.

Its 5-parameter analogue is, however, satisfactory for water. The results obtained on the basis of the eq 3-8 and 10a are compared in Table I. The constants A , B , C, and D and the exponents f, g, and h of the Cparameter eq 3, 5,7, and 9 are recorded in Table I I. The constants A , B , C , D , and E and the exponents f , g , h , and iof the 5-parameter equations are recorded in Table 111. The exponents f, g , h, and i o f eq 8 are from Wagner and Polak (23).The ESDU (3)has chosen the values 1.5, 3.0, and 6.0 for f, g , and h of the 4-parameter Wagner equations. Recently, Wagner, Ewers, and Penterman (24) have suggested the values 1.5, 2.5,and 5.0,but the values 1.5, 2.25, and 4.25 found by us fit the vapor pressure data of water adequately. We have also recorded in Tables I 1 and 111 the a values obtained on the basis of all these equations at T,, T, and T,. The value of (3 below the normal boiling point and the standard deviation, u, is also recorded in Tables I1 and 111. From a survey of the results presented in Table I, it can be seen that the most satisfactory of all the equations is the 5parameter eq 4b. The eq 3b containing only 4 parameters is slightly less accurate than eq 4b. The Wagner equation (7) as well as the new equation (eq 5)do not fit the vapor pressure around the normal boiling point satisfactorily. Judging from the a value at the triple point, Wagner equation (7) and the new equation (5)are not satisfactory at the triple point. The best values for at and abmay be obtained from the cubic equation. The constants of eq 1 and 2 are recorded in Table I V . They have been obtained by using the vapor pressure data of water in the range 270-373.15 K. The vapor pressure equations proposed by others for water ( 75,25,26,27)lack the simplicity of the equations considered in this article. Also the equation proposed by Scott and Osborn ( 15)is an analytic one and violates the criterion 1. According to one of the referees of this article, the defect of any equation in the region below the normal boiling temperature is undetectable on the basis of the standard deviations calculated by using the absolute deviations. He recommended

Journal of Chemical and Englneerlng Data, Vol. 31, No. 4, 1986 441

Table I. Comparison of the Vapor Pressure Equations for Water (P(obsd) - P(calcd))/kPa 1D.O

T/K

P/kPab

(3a)

(3b)

(5)

(7)

2 1 2 1 5 6 2 2 2 2 2 1 2 2 2 1 2 2

273.150 273.150 273.150 273.160 273.160 273.160 273.160 274.150 275.150 276.150 277.150 278.150 278.150 280.650 280.650 283.150 283.150 285.650 288.150 288.150 290.650 293.150 293.150 298.150 298.150 313.150 323.150 333.150 343.150 353.150 373.150 383.154 393.157 403.162 413.166 432.170 433.175 443.179 453.184 463.188 473.193 483.197 493.201 503.204 513.208 523.211 533.214 543.217 548.218 553.219 563.221 573.223 583.224 593.225 598.226 603.226 613.227 623.227 628.227 633.227 635.227 637.227 639.227 641.227 643.227 644.227 645.227 646.227 647.126

0.509 .~ 0.611 0.611 0.611 0.612 0.612 0.612 0.657 0.706 0.759 0.814 0.872 0.873 1.037 1.037 1.228 1.229 1.449 1.705 1.707 2.001 2.339 2.340 3.169 3.169 7.381 12.345 19.933 31.177 47.375 101.325 143.304 198.577 270.112 361.335 475.913 618.062 792.139 1002.692 1255.285 1555.055 1907.950 2319.978 2797.756 3347.940 3977.675 4694.245 5505.322 5949.044 6419.526 7445.148 8592.046 9869.683 11289.429 12056.459 12864.222 14607.721 16536.747 17577.152 18673.894 19128.843 19594.026 20068.835 20554.486 21052.397 21305.608 21561.453 21821.656 22060.000

0.000 -0.001 0.000 0.000 0.000 0.000

0.000

0.000

0.000

-0.001 0.000 -0.001

-0.001

-0.001 0.000 0.000 0.000

1 2 2 1 2 1 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 7

~

0.000 0.000 0.000

0.000 0.000 0.000 0.000 0.000

0.001 0.000 0.000 0.000 0.000 0.000 0.000

0.001

0.001 -0.001 0.000 0.001

0.001 -0.001 0.000 0.001 0.001

0.000 0.000 0.000 0.000 -0.001 -0.001 -0.001 -0.001 0.000

0.002 0.007 0.049 0.056 -0.002 -0.028 -0.078 -0.035 0.046 -0.117 -0.050 -0.133 -0.151 -0.234 -0.144 -0.074 0.145 0.307 0.257 0.380 0.455 0.355 0.453 0.234 -0.117 -0.357 -0.034 -0.074 -0.241 -0.667 -0.038 0.033 0.442 0.161 -0.114 0.341 0.192 -0.663 -0.835 -0.429

0.000 0.000 0.000 0.000 0,000 0.000

0.000 0.001 0.000

0.000 -0.001 -0.003 -0.005 -0.007 -0.008 0.000 0.054 0.082 0.056 0.074 0.076 0.176 0.310 0.186 0.264 0.152 0.055 -0.163 -0.264 -0.429 -0.467 -0.554 -0.799 -0.739 -0.696 -0.747 -0.425 -0.246 -0.074 -0.038 0.543 0.862 0.650 0.006 0.324 0.271 0.580 0.253 0.032 0.721 0.802 0.296 0.664 1.905

0.000 0.000 0.000

0.000 0.001 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000 -0.001 0.001 -0.001 -0.001 0.001 0.000 -0.001 0.000 -0.001 -0.002 -0.003 -0.003 -0.002 0.001 0.007 0.028 0.082 0.102 0.056 0.039 -0.008 0.030 0.094 -0.098 -0.075 -0.215 -0.299 -0.452 -0.425 -0.399 -0.194 -0.004 0.028 0.213 0.362 -0.447 0.756 0.740 0.532 0.323 0.639 0.456 -0.022 -0.633 -0.158 -0.127 0.265 -0.003 -0.224 0.337 0.267 -0.490 -0.540 0.000

0.000 0.000 0.000 0.000 0.001

0.000 0.000 0.000 0.000 0.000 0.000 0.001 -0.001 -0.001 0.001 0.000 -0.001 0.000 -0.001 -0.001 -0.001 -0.001

0.001 0.005 0.010 0.024 0.069 0.076 0.014 -0.023 -0.091 -0.073 -0.024 -0.220 -0.185 -0.294 -0.322 -0.394 -0.267 -0.130 0.186 0.465 0.543 0.726 0.854 0.829 0.931 0.617 0.058 -0.327 -0.168 -0.550 -0.948 -1.374 -0.616 -0.449 0.084 -0.045 -0.143 0.505 0.452 -0.315 -0.419 0.000

(44 0.000 -0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000

(4b) 0.000 -0.001

0.001 0.000

(6) 0.000

(8)

0.000

0.000 0.000

-0.001 0.000 0.000

0.000 0.000 0.000 0.000 0.000

0.000 0.000 0.000 0.000 0.000

0.000 0.000 0.000 0.000 0.000 0.000

0.001 0.000 0.000 0.000 0.000

0.001 0.000 0.000

0.001 0.000 0.000

0.000 0.000

0.000 0.000 0.000 0.000

0.000 0.000 0.000 0.000

0.001 0.000

0.001 -0.001 0.000 0.001 0.000

0.001 -0.001 0.000 0.001 0.000

0.001 -0.001 0.000 0.001 0.001

0.000

0.000 0.001 0.000 -0.001 -0.001 -0.001 -0.001 -0.002 -0.001 0.000 0.042 0.048 -0.008 -0.030 -0.074 -0.021 0.071 -0.078 0.002 -0.072 -0.087 -0.174 -0.100 -0.056 0.127 0.247 0.154 0.258 0.316 0.196 0.294 0.097 -0.216 -0.437 -0.100 -0.141 -0.370 -0.843 -0.238 -0.158 0.283 0.068 -0.089 0.568 0.569 -0.088 0.009 0.777

0.000 0.001 0.000

0.000

0.001 0.000

0.000 0.000 0.000 0.000 0.000 0.001

-0.001 0.000

-0.001 -0.001 -0.001 -0.002 -0.002 -0.001 0.001 0.044 0.051 -0.005 -0.027 -0.072 -0.021 0.069 -0.082 -0.004 -0.078 -0.091 -0.174 -0.093 -0.042 0.148 0.270 0.173 0.270 0.319 0.169 0.227 -0.013 -0.353 -0.574 -0.221 -0.167 -0.208 -0.568 0.120 0.209 0.631 0.357 0.080 0.521 0.259 -0.514 -0.712 -0.338

-0.001 -0.001 -0.003

-0.004 -0.005 -0.005 0.000 0.047 0.061 0.016 0.005 -0.027 0.033 0.125 -0.035 0.020 -0.094 -0.163 -0.317 -0.311 -0.330 -0.190 -0.081 -0.138 0.002 0.111 0.133 0.420 0.437 0.333 0.204 0.618 0.673 0.446 -0.067 0.458 0.489 0.862 0.555 0.266 0.723 0.582 -0.259 -0.415 0.000

-0.001 0.000

0.001 0.000 -0.001 -0.001 -0.003 -0.004 -0.006 -0.007 -0.004 0.043 0.059 0.016 0.010 -0.017 0.050 0.149 -0.005 0.053 -0.063 -0.140 -0.308 -0.326 -0.375 -0.269 -0.193 -0.276 -0.142 -0.036 0.000 0.331 0.427 0.432 0.363 0.838 0.995 0.798 0.254 0.706 0.696 1.022 0.662 0.320 0.727 0.567 -0.287 -0.442 0.000

(loa) 0.000

(3b)' 0.000

-0.001 0.000 0.000

-0.001 0.000 0.000 0.000

0.000 0.000 0.000 0.000 0.000 0.001

0.000 0.000 0.000 0.000 0.000 0.000 0.001 -0.001 0.000 0.001 0.000 0.000 0.001 0.000

-0.001 -0.001 -0.002 -0.001 -0.001 0.000 0.005 0.048 0.056 -0.001 -0.027 -0.078 -0.035 0.043 -0.121 -0.054 -0.135 -0.145 -0.214 -0.103 -0.006 0.244 0.433 0.396 0.515 0.577 0.416 0.395 -0.005 -0.577 -0.928 -0.701 -0.818 -0.764 -0.914 0.098 0.337 0.905 0.752 0.533 0.895 0.598 -0.519 -1.150 -1.564

0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.001 -0.001 0.000 0.001 -0.001 0.000 0.001 0.000 -0.001 -0.002 -0.004 -0.007 -0.010 -0.011 0.000 0.058 0.093 0.078 0.109 0.130 0.252 0.411 0.313 0.418 0.330 0.251 0.041 -0.067 -0.259 -0.350 -0.517 -0.876 -0.886 -0.923 -1.156 -1.046 -1.099 -1.159 -1.232 -0.747 -0.561 -0.765 -1.320 -0.833 -0.791 -0.367 -0.558 -0.617 0.263 0.453 0.067 0.566 1.940

OID. 1, Besley and Bottomley (16); 2, Douslin (17); 3, Stimson (18); 4, Osborne, Stimson, Fiock and Ginnings (19); 5, Prytz (20); 6, Johnson, Guildner, and Jones (21); 7, NBS/NRS Steam Tables for critical temperature (22). *For the sake of this table the observed pressures were shown rounded to three places after the decimal for the first 30 observations. This refers to the deviations obtained on the basis of eq 3b with the points above the normal boiling point and below the critical point not participating in the regression. In place of all these points, an additional point with 2' = 580.000 K and pressure equal to 9442.527 kPa calculated from eq 21 was used.

442 Journal of Chemical and Engineering Data, Vol. 3 7,

No. 4, 7986

Table 11. Constants of t h e 4-Parameter Eauations for Water equation const

A B

c

D

f g

3a -17.678684 0.828603 -3.314618 -4.312329 1.50 3.29

3b -11.868212 1.999582 -4.451665 -8.025567 1.55 3.95

5 -7.862038 -0.2031 24 1.917098 -5.136087 1.50 6.00

0.225 0.00030 0.00045 19.846 13.316 7.867 0.018

0.338 0.00034 0.00050 19.829 13.312 7.820 0.018

0.266 0.00034 0.00049 19.859 13.315 7.862 0.018

h U

u1(1 Orb

at ab

P

7 -7.869123 2.003247 -1.813041 -2.405901 1.50 2.25 4.25 0.372 0.00032 0.00046 19.854 13.316 7.969 0.018

9a 8.809532 2.753596 -1.069863 2.649319 1.32

9b 4.464503 1.083562 -0.164437 1.171406 1.22

1.208 0.00036 0.00053 19.829 13.312 8.113 0.018

1.624 0.00037 0.00051 19.825 13.312 8.360 0.018

"Standard deviation based on relative deviations from the triple point to the critical point. bStandard deviation based on relative deviations below the normal boiling temperature. Table 111. Constants of the 5-Parameter Equations for Water const A

B C D E

f

i?

4a -17.650073 0.921524 0.075143 -3.332385 -4.501325 1.50 3.32

4b -11.880951 1.817797 -0.358975 -4.129636 -6.009070 1.50 3.50

eauation 6 -7.867226 -0.235995 0.031495 1.947708 -7.650007 1.50 6.50

0.222 0.00030 0.00046 19.843 13.315 7.867 0.018

0.196 0.00030 0.00046 19.844 13.315 7.820 0.018

0.272 0.00031 0.00047 19.840 13.314 7.862 0.018

h

i U

0,"

at

ab

a,

P

8 -7.864939 1.902006 -2.379234 -9.493641 9.356228 1.50 2.50 6.50 7.50 0.340 0.00031 0.00047 19.839 13.314 7.869 0.018

Standard deviation based on relative deviations from the triple point to the critical point. ations below the normal boiling temperature.

10a 8.385026 3.652543 -1.333139 -0.211373 3.081890 1.45

10b 4.585420 1.097441 -0.279857 -0.019689 1.051922 1.37

0.393 0.00030 0.00046 19.843 13.315 7.922 0.018

0.485 0.00030 0.00046 19.843 13.315 7.984 0.018

Standard deviation based on relative devi-

the calculation of the standard deviations based on the relative deviations given by (Pexptl - P,,,)/P exptl. We have, therefore,

calculated such standard deviations based on the relative deviations and recorded them in the Tables 11-XIII, using the symbol gr. These standard deviations do not, however, affect our earlier conclusions but they help highlight the defects of the vapor pressure equations below the normal boiling temperature while they mask the large absolute deviations at higher temperatures. I n any case we agree with the referee that the standard deviations based on relative deviations are more meaningful than the standard deviations based on absolute deviations. The relative deviations obtained on the basis of the 4-parameter equations are shown graphically in Figures 3 and 4. The above equations have been tested for a wide variety of substances and the same conclusions are drawn as with water.

Table V. Constants of t h e Eq 3a for Various Substances const methane benzene A -10.182850 -13.349312 B 2.304823 0.781938 -2.201223 C -3.122045 D -5.375955 -6.637640 1.50 1.50 f 3.35 3.35 190.555 562.161 90.680 353.242 U 0.284 0.290 0.00020 0.00022 or 11.636 15.137 fft 9.196 10.873 ffb 5.995 7.009 0.089 0.022

nitrogen -10.208219 2.015495 -2.392160 -5.540867 1.50 3.35 126.200 63.148 0.189 0.00028 11.595 9.121 6.116 0.127

Table IV. Constants of t h e Eq 1 a n d 2 for Water eauation lb 2a 3a const la A -19.896537 -13.363884 -19.847177 -13.314231 B -1.643660 -2.516920 -1.969654 -3.066515 273.168660 373.119095 273.160517 373.149768 c 1.441277 0.565395 D 0.001 U 0.001 0.023 0.023 0.00038 0.00038 0.00076 0.00076 or 19.847 19.847 19.897 19.897 fft 13.314 13.314 13.364 13.364 0.018 0.018 0.018 0.018

argon -8.572984 2.290292 -2.352673 -5.856691 1.50 3.35 150.690 83.804 0.371 0.00029 9.654 9.220 5.898 0.110

oxygen -14.937528 1.812160 -2.087957 -4.522111 1.50 3.35 154.581 54.361 0.164 0.00011 17.086 9.438 6.079 0.110

Journal of Chemical and Engineering Data, Vol. 3 1, No. 4, 1986 443 A

LEGEND 0.0004

E

o EQN. ( 3 a )

0

-0 . 0 0 0 3 -0.0004

A

EON. ( 3 b )

A

t

A

I

I

I

I

I

Table VI. Constants of the Eq 3b for Various Substances const

A B C

D f g TC Tb U

ur fft ffb ffC

R

methane -8.170039 2.518205 -3.056869 -7.893296 1.55 3.95 190.555 111.632 0.264 0.00017 11.640 9.192 6.008 0.089

benzene nitrogen -9.732746 -8.196031 1.891007 2.272766 -4.334186 -3.155023 -10.219543 -8.357627 1.55 1.55 3.95 3.95 126.200 562.161 353.242 77.348 0.250 0.301 0.00022 0.00022 11.621 15.144 9.117 10.872 6.102 6.999 0.126 0.022

argon -8.210497 2.356966 -2.850425 -8.052901 1.55 3.95 150.690 87.290 0.423 0.00026 9.662 9.219 5.905 0.110

oxygen -8.432226 2.274787 -2.890493 -8.372496 1.55 3.95 154.581 90.188 0.253 0.00020 16.925 9.439 6.025 0.111

The results on the basis of eq 3 and 4 for a few typical compounds are shown in Tables V-VIII. The compounds include nitrogen, argon, oxygen, methane, and benzene. The vapor pressure data for these compounds are from reliable sources (1, 15, 24, 28-34). Judging from the standard deviations recorded in Tables V-VIII, we find these equations to be extremely satisfactory for these compounds. We have also found it possible to fix the values of the exponents f and g for eq 3a, 3b, 4a, and 4b for practically all the substances tested in this

I

I

I

I

I

I

I

I

article. Since the triple points and the low-pressure data are usually not available, we select eq 3b and 4b.

Predictive Procedures 1 . Zeroth-Order Approxbnation. We have already shown (4) how the quadratic equation can be used for the prediction of vapor pressure from the triple point to the normal boiling point. A simple equation such as eq 21 can now be used to obtain the vapor pressure from the normal boiling point to the critical point as shown below. Equation 21 is referred to as the Clapeyron equation and is also discussed by Reid, Prausnitz, and Sherwood (35). In P = a

- b/T

(21)

where b = AHIAZR. On the basis of eq 13 and 14 CY is given by the following expression: CY

= AH/AZRT

(22)

Equation 21 may, therefore, be written as follows:

(23)

InP=a-a

I t follows from eq 23 that when the pressure is expressed as P I f or P I f b, or P I f c, the following conditions are obeyed

,,

Table VII. Constants of the Eq 4a for Various Substances const

A B C D E

f g

Tc Tt U

0,

at "b ffC

4

methane -10.277025 1.875931 -0.477635 -2.265333 -4.453736 1.50 3.35 190.555 90.680 0.097 0.00013 11.648 9.193 6.014 0.089

benzene -13.391401 0.571577 -0.252950 -3.165810 -6.111294 1.50 3.35 562.161 278.68 0.255 0.00021 15.142 10.873 7.020 0.022

nitrogen -10.198031 2.058812 0.045962 -2.391777 -5.634290 1.50 3.35 126.200 63.148 0.192 0.00028 11.594 9.121 6.115 0.127

argon -8.883885 0.873589 1.614501 -2.233650 -1.922376 1.50 3.35 150.651 83.804 0.375 0.00024 9.671 9.220 5.903 0.109

oxygen -15.104085 1.275208 -0.432193 -2.096072 -4.085843 1.50 3.35 154.581 54.361 0.203 0.00009 17.155 9.433 6.085 0.109

444 Journal of Chemical and Engineering Data, Vol. 3 1, No. 4, 1986 Table VIII. Constants of the Eq 4 for const methane benzene nitrogen A -8.250120 -9.801688 -8.263900 B 2.139985 1.499906 1.914219 C -0.442781 -0.549661 -0.471124 D -2.641107 -3.807586 -2.744418 E -5.231639 -6.674194 -5.374371 f 1.50 1.50 1.50 g 3.50 3.50 3.50 TC 190.555 562.161 126.200 Tb 111.632 353.242 77.348 0.268 u 0.094 0.187 ur 0.00013 0.00021 0.00021 at 11.647 15.152 11.624 ab 9.193 10.873 9.119 6.005 7.003 6.103 0.089 0.022 0.126

ic

respectively: (1) a = at at T,; (2) a =

and the critical temperature. We shall refer to this procedure as the zeroth-order approximation. The a values which we obtain on the basis of the eq 26 differ slightly from those which we obtain from the eq 1-10. We may resort to this procedure when no vapor pressure data are available above the normal boiling temperature. 2 . First-Order Approximation. We shall now describe the first-order approximation to the vapor pressure. I t consists in using all the available vapor pressure data up to the normal boiling point and then just two additional data points above the normal boiling temperature. These two additional points are (1) the T , and P , and (2) Ti and Pi. For testing the above hypotheses and also for testing which of the ten equations considered in this paper would be most appropriate for predictive purposes, we use water as our example. The TI for water is close to 580 K ( T , = 0.90) as can be seen from the Figure 1. The value of a = 12.715 and of b = -4744.597 for water. The pressure at Ti calculated from eq 13 is 9442.527 kPa. The constants as well as the standard deviations, ds, and nonweighted standard deviations of all the equations are shown in Tables I X and X. The nonweighted standard deviation is the standard deviation obtained on the basis of deviations of all the points used and unused in the regressions. These equations fit the low-pressure data with sufficient accuracy and then predict the vapor pressure above the normal boiling point quite satisfactorily. The fit is excellent with the eq 3b. The results obtained on the basis of eq 3b based on the predictive procedure are also shown in Table I.They compare favorably with the deviations obtained on the basis of the regular procedures. The results obtained on the basis of the 5-parameter equations are not satisfactory, presumably due to overfitting as pointed by one of the referees. The question now is as to how we can obtain the four values namely, T,, P,, Ti, and P , for substances for which they are not known. Fortunately, several empirical procedures (36-44) are now available for the prediction of both T , and P,. Equations 11, 12, and 26 may also be used for the estimation of the critical pressure. Several deviation plots for the isomeric hexanes were drawn by Kay (45).McMicking and Kay ( 4 6 ) listed the calculated deviations for heptanes and octanes. From these we see that TI occurs normally around a T , of 0.88. The T , is slightly above the temperature at which AHlAZ goes to a minimum. According to McGarry (47), the temperature at which the minimum in AHlAZ occurs is a function of the normal boiling point: the higher the normal boiling point, the higher this temperature would be. Similarly, T , also may be considered to be a function of the normal boiling point. For the purpose of the first-order approximation, we shall assume that T , occurs around T , = 0.90. We shall also assume that the exponents f and g are optimized at 1.55 and 3.95 for eq 3b. The only unknown now is Pi. I n order to calculate P , from eq 26, we shall introduce a factor called the flex factor

Various Substances argon oxygen -8.4456922 -8.479508 1.229753 1.929615 -1.348862 -0.558801 -2.383031 -2.695696 -2.536030 -4.847068 1.50 1.50 3.50 3.50 150.690 154.810 87.290 90.188 0.389 0.124 0.00024 0.00008 9.668 17.241 9.219 9.430 5.907 6.055 0.109 0.109 ab

at Tb; (3) a = a, at

TC.

I f we constrain eq 21 to go through the normal boiling point and the critical point, and express the pressures in atmospheres, the constants a and b of the eq 21 will be given by the following expressions:

b = aTb Equation 21 may now be simplified as follows

In P = -axb

(24)

In ( P / P , ) = -aYTb/T,

(25)

which may be written alternatively as follows: In P =

(26)

In (P/P,) = -a,Y

(27)

Figure 1 illustrates the type of deviations one obtains by the use of eq 26 for water, argon, and ethanol. Around T , = 0.90 the deviations change their sign for water and several classes of substances such as hydrocarbons, esters, ketones, etc. These compounds have an inversion point, Ti. Approximately around the same T, compounds such as alcohols, amines, etc., have a maximum positive deviation. I n such cases this point is referred to as the inflection point. Compounds such as argon and halogenated hydrocarbons, on the other hand, have maximum negative deviation close to this T,. These deviations, on a relative basis, are, in general, so small for most of these compounds that one may use eq 26 itself for the calculation of the vapor pressure between the normal boiling temperature Table

IX. Constants of the Predictive 4-Parameter Equations for Water ~

equation const A

B C D

f g

3a -17.625665 0.919542 -3.441608 -4.196342 1.50 3.29

3b -11.876514 1.979964 -4.415632 -8.026502 1.55 3.95

-7.900076 -0.248706 2.068462 -3.228613 1.50 6.00

0.00068 0.00065 19.828 13.322 7.888 0.018

0.00049 0.00021 19.833 13.311 7.815 0.018

0.00048 0.00096 19.860 13.322 7.900 0.018

5

h u/ Urb

at ab "C

P

7 -7.886539 2.076640 -1.913650 -2.332365 1.50 2.25 4.25 0.00044 0.00053 19.849 13.320 7.887 0.018

8.798717 2.773195 -1.078937 2.653968 1.32

9b 4.477823 1.066982 -0.161704 1.157570 1.22

0.00067 0.00033 19.817 13.312 8.113 0.018

0.00050 0.00038 19.831 13.311 8.344 0.018

9a

Standard deviation based on relative deviations below the normal boiling temperature. 'Standard deviation based on relative deviations for all observations above the normal boiling temperature.

Journal of Chemical and Englneering Data, Vol. 3 1, No. 4, 1986 445

Table X. Constants of the Predictive 5-Parameter Equations for Water equation const

A B C

D E

f g h

4a -17.550273 1.356613 0.530766 -3.214357 -5.303798 1.50 3.32

4b -11.867620 1.817299 -0.435004 -4.246483 -5.679914 1.50 3.50

6 -7.897848 -0.263622 0.021672 2.065603 -5.281576 1.50 6.50

0.00068 0.00061 19.820 13.312 7.845 0.018

0.00045 0.00036 19.848 13.318 7.865 0.018

0.00045 0.00081 19.847 13.320 7.898 0.018

i 0:

6 :

at ab

a.2

P

8 -7.906216 2.056907 -2.635651 -6.781610 5.938759 1.50 2.50 6.50 7.50 0.00045 0.00097 19.847 13.320 7.906 0.018

10a 8.525269 3.499316 -1.413453 -0.118149 2.928770 1.45

10b 4.550555 1.148475 -0.291466 -0.024374 1.087985 1.37

0.00068 0.00036 19.819 13.313 7.896 0.018

0.00045 0.00036 19.848 13.318 8.002 0.018

Standard deviation based on relative deviations below the normal boiling temperature. *Standard deviation based on relative deviations for all observations above the normal boiling temperature.

Table XI. Constants of the Ea 26 and 27 and the Flex Factors for Various Substances compound water methane ethane propane butane ethylene acetylene cyclohexane benzene methanol ethanol 1-propanol 2-propanol 1-butanol 2-butanol 2-methyl-1-propanol 2-methyl-2-propanol 1-pentanol 1-octanol chlorine ethyl fluoride acetone acetic acid ethyl, acetate diethyl ether 1-propylamine 2-propylamine nitrogen oxygen argon carbon monoxide carbonyl sulfide

and denote it by the symbol stance is defined as

x

ab

12.715 9.212 9.791 9.954 10.094 9.768 10.558 10.239 10.436 12.837 12.986 12.679 12.800 12.342 12.212 12.309 12.322 12.138 11.836 10.242 10.477 10.899 11.888 11.025 10.501 10.853 10.755 9.076 9.380 9.202 9.184 10.054

x.

The flex factor,

= PMexpt~)/Pycal~d)

T,

a C

Tb

7.332 5.397 5.918 6.219 6.473 5.864 6.454 6.545 6.557 8.457 8.880 8.747 8.950 8.568 8.490 8.562 8.653 8.485 8.495 5.876 6.573 7.062 7.843 7.380 6.920 6.996 6.951 5.563 5.473 5.330 5.644 5.916

373.150 111.632 184.554 231.054 272.66 169.415 188.472 353.881 353.242 337.696 351.440 370.300 355.390 390.882 372.660 381.04 355.49 411.15 468.35 239.184 235.45 329.217 391.035 350.26 307.581 320.369 304.907 77.348 90.188 87.290 81.638 222.9

x for a sub-

at Tr = 0.90

(28)

where PyceM) is obtained from eq 26. The flex factor may be considered as a measure of the complexity of a molecule at T, = 0.90. I t may also be related to the van der Waals constant a . We assume that the pressure is normal when it is given by the eq 26. This is the case when x = 1. When < 1, we assume that the pressure is lessened because of attraction by the mass of molecules in the bulk gas. When x > 1 the pressure is enhanced on account of repulsion. The flex factor may alternatively be looked upon as the power of a molecule to attract other molecules towards itself. Thus it may find great use In the study of mixtures. The flex factor may be related to Pitzer’s acentric factor ( 4 8 ) as well as to Stiel’s polarity factor (35)in concept, but they represent different functions. These factors, in general, represent different func-

x

647.126 190.555 305.33 369.85 425.14 282.2 308.32 553.64 562.161 512.640 513.920 536.775 508.296 563.051 536.015 547.778 506.205 588.15 652.5 416.90 375.31 508.10 592.71 523.30 466.74 497.0 471.8 126.20 154.581 150.69 132.85 378.8

X 1.000 0.974 0.984 0.989 0.992 0.986 0.998 0.996 0.997 1.014 1.044 1.067 1.070 1.070 1.070 1.075 1.075 1.050 1.009 0.990 1.000 0.998 0.998 1.005 1.002 1.015 1.015 0.979 0.977 0.976 0.983 0.987

ref

tions at different T,. The flex factors of some molecules are recorded in Table X I together with the constants of eq 26 and 27. Given the flex factor, one can calculate P , and use the first-order approximation to the vapor pressure. Fortunately, for a large number of substances, the flex factor is equal to unity and we encounter relatively little problem in the prediction of vapor pressure. For substances for which the flex factor is not equal to unity, it is necessary to estimate it. For example, if we know the flex factor for I-pentanol, we can assume the same factor to be valid for all of its isomers. Thus we can, in general, use the first-order approximation to the vapor pressure. Test of this procedure for a large number of substances has revealed that the procedure is, indeed, quite satisfactory. Similar results are obtained with eq 3a with the exponents f and g having the values 1.5 and 3.35. We have recorded the results obtained with eq 3b in Table XI1 for a number of substances. The vapor pressure data for these compounds are taken from reliable sources ( 4 9 - 7 7 ) . We have used graphical interpola-

446 Journal of Chemical and Engineering Data, Vol. 3 1, No. 4, 1986

Table XII. First-Order Approximation: Constants of the Ea 3b for Various Substances A compound B C D water methane ethane propane butane ethylene acetylene cyclohexane benzene methanol ethanol 1-propanol 2-propanol 1-butanol 2-butanol 2-methyl-1-propanol 2-methyl-2-propanol 1-pentanol 1-octanol chlorine ethyl fluoride acetone acetic acid ethyl acetate diethyl ether 1-propylamine 2-propylamine nitrogen oxygen argon carbon monoxide carbonyl sulfide

-11.908768 -8.196 404 -9.462 159 -8.952 340 -9.007 002 -8.994 073 -10.045 373 -9.527 448 -9.717 584 -12.727971 -13.271 159 -13.125612 -13.250821 -12.655 908 -12.305 436 -12.699 812 -12.503 713 -11.651 089 -10.042 394 -9.477 864 -10.073 233 -10.302 747 -10.982 406 -10.443 097 -9.942 502 -10.482 203 -10.243 682 -8.166 938 -8.405 606 -8.029 280 -8.332 804 -9.122 383

1.906 813 2.452 845 0.798 532 2.655 006 3.049 617 1.935 515 1.319277 1.971 018 1.932 356 0.320 894 -0.902 147 -1.439 823 -1.498 131 -1.025 719 -0.308 134 -1.282989 -0.610 480 1.616401 -6.370 723 1.792 787 1.085 470 1.692518 2.660 199 1.761 144 1.636 966 1.042 553 1.449 188 2.347 970 2.340 113 2.787 231 2.208 147 2.260 928

-4.317 389 -2.960 683 -1.411 911 -5.505 578 -5.873 902 -3.057 512 -2.905 940 -4.453 953 -4.395 194 -2.658 842 -3.951 898 -5.975 239 -6.672 664 -7.705 747 -9.709805 -7.798 184 -10.007 340 -12.177628 -18.616 030 -3.038 638 -2.616 926 -4.028711 -4.182 332 -5.286084 -4.477 357 -4.467469 -5.396020 -3.279657 -2.985 541 -3.924 632 -3.399 466 -3.959088

-7.886 524 -7.816 273 -5.957 056 -6.363 559 -9.651 789 -8.477 172 -9.799 399 -10.164 153 -10.296 149 -3.394 946 5.190 496 8.733 184 10.615 046 3.495 925 0.323 374 4.569 191 4.309 713 -5.842219 -27.682 396 -7.384 808 -5.011 352 -9.096 402 -14.367 847 -10.159 110 -8.845 618 -10.932 741 -9.480 064 -8.405 931 -8.470 138 -5.492 141 -8.377 530 -7.995 762

~~~~~

: 0

0.000 52 0.000 22 0.003 75 0.006 55 0.002 70 0.000 76 0.000 40 0.000 64 0.000 15 0.000 30 0.000 13 0.000 09 0.000 09 0.000 05 0.000 08 0.000 19 0.000 07 0.000 27 0.002 86 0.000 70 0.003 59 0.000 20 0.000 52 0.000 21 0.000 35 0.000 05 0.001 31 0.000 45 0.000 18 0.000 24 0.001 13 0.002 19

0 :

0.000 58 0.000 24 0.004 84 0.00581 0.003 29 0.000 47 0.002 37 0.000 59 0.000 30 0.002 00 0.001 31 0.002 26 0.002 69 0.004 20 0.011 48 0.003 05 0.002 27 0.001 42 0.002 01 0.004 07 0.006 33 0.002 09 0.000 37 0.001 76 0.002 35 0.001 23 0.002 49 0.000 34 0.000 26 0.003 92 0.000 57 0.005 60

a Relative standard deviation for weighted observations below the normal boiling temperatures. *Relative standard deviation for all observations above the normal boiling temperature.

Table XIII. Mathematical Expressions for the a-Function Based on Eq 1-10” ea

a

(3) - ( T i ) n i A + 2BX + CZf +‘DZg] - (X/TJ[fCZf’ + gDZg’] (4) - ( T x / T ) [ A+ 2EX + 3CX2 DZf + EZg] - (X/T,)[fDZf + gEZg‘] (5) - W [ A + 2BYI - (l/T,)[fCZf + gDZg] (6) -W[A + 2BY + 3 C p ] - (l/Tc)[fDZf + gEZp‘1 (7) - W [ A + BZf + CZg + DZh] - [ f B Z f+ gCZf + hDZh] (8) - W [ A + BZf + CZg + DZh + EZ’] - [fEZf + gCZg’ + hDZh’ + iEZ”] (9) - ( T , / T G ) [ A+ 2BX + 3CX2 + D(Y/H)f] ( TcX/TGH)[fD( Y/iWf1 (10) -(T,/TG)[A + 2BX + 3CX2 + 4DX3 + E ( Y / M f ] (TJ/TGH)[ f E (y/Wf1

+

In the above expressions we have f‘ = f - 1;g’ = g - 1; h’ = h 1; i’ = i - 1. From a one can derive the following: d P / d T = a P / T AHJAZ = ( R P / P ) ( d P / d T ) = RTa. tions to obtain the flex factors of some of the above molecules. 3 . Second-order Appmxhation. A third procedure, which may be referred to as the second-order approximation to the vapor pressure, consists in using one actual experimental data point closest to T,. The results obtained by using this procedure are extremely satisfactory for a number of compounds and are essentially similar to those already shown in Table XII. The advantage in this procedure is that it eliminates the need to determine vapor pressures in the entire range from the normal boiling point to the critical point. 4 . The Group -Additlvffy Procedures. Other predictive procedures involve prediction of the constants of the vapor pressure equations. Several group additivity procedures have been described by us previously (4) for the estimation of the constants of the quadratic equation with respect to alkanes. Similar group additivity procedures may also be used for the

estimation of the constants of some of the vapor pressure equations considered in this paper. For example, for the calculation of the value cyc of the alkanes we have the following group additivity equation

a, = an2’3 -I-a , n ,

+ a 2 n 2+ a 3 n 3 + a 4 n 4

(29)

where n is the number of carbon atoms, n is the number of primary carbon atoms, n 2 is the number of secondary carbon atoms, n3 is the number of tertiary carbon atoms, and n 4 is the number of quaternary carbon atoms in the alkane molecule. The constants have the following values: a = 0.22278; a = 3.04664; a 2 = 0.15871; a 3 = -2.85149; a 4 = -5.97162. These constants are obtained on the basis of the Wagner constant A reported by ESDU (3) for 80 alkanes in the range C,-C,,. Such group additivity procedures will be considered elsewhere after we apply the vapor pressure equations to various classes of compounds.

,

Conclusion I n conclusion, we wish to remark, that it is no longer desirable to make extensive compilations of the Antoine or other type of constants based on low-pressure data. The predictive procedures shown in this paper produce a good fit for the low-pressure data and also provide vapor pressures above the normal boiling point up to the critical point. We find the firstorder approximation using the flex factors to be quite simple to use and is very accurate in the prediction of vapor pressure above the normal boiling temperature. We also wish to make the general observation that none of the eparameter equations may be accurate at all the three points, namely, T,, T, and T., For a better representation of the vapor pressure data from the triple point to the critical point one should resort to the 5-parameter equations such as eq 4a or eq 4b. This may be necessary only when accurate vapor pressures are available.

Journal of Chemical and Engineering Data, Voi. 31, No. 4, 1986 447

Goodwin ( 4 9 , 50) has also emphasized the cubic behavior of the residuals of eq 21. Acknowledgment My thanks are due to Dr. K. N. Marsh and Dr. R. C. Wilhoit for their kind interest in this work. My appreciation is due to Ramgopal V. Gollakota, Mary Jane Rodriguez, Linda Ryan, and Mark Sutton for help with the preparation of the manuscript. M y special thanks are due to all the referees of this article for their helpful comments. Literature Clted (1) Wagner, W. Cryogenics 1973, 70. 470. (2) Ambrose, D. J. Chem. Thermodyn. 1978, IO, 765. (3) Engineering Sciences Data Unit, ESDU International Ltd., London, 1985. (4) Somayajulu, G. R. Hlgh Temp. Scl. 1984, 77, 427. (5) Goodwin, R. D. J. Res. Natl. Bur. Stand. 1989, 73A, 487. (6)Levetl Sengers, J. M. H.; Chen, W. T. J. Chem. Phys. 1972, 56,595. (7) Fisher, M. E. J. Math. Phys. 1984, 5, 944. (8) Yang. C. N. Phys. Rev. Lett. 1964, 73,303. (9) Warlng, W. Ind. Eng. Chem. 1954, 4 6 , 762. (10) Watson, K. M. Ind. Eng. Chem. 1943, 35, 398. (11) Gambill, W. R. Chem. Eng. 1957, 6 4 , 261. (12) Ambrose, D.; Counseli. J. F.; Hicks, C. P. J. Chem. Thermodyn. 1978, 70, 771. (13) Plank, R.; Riedel, L. 1 s . - A r c h . 1948, 16. 255. (14) Ambrose, D.; Davles, R. H. J. Chem. Thermodyn. 1980, 72, 871. (15) Scott, D. W.; Osborn, A. 0. J. Phys. Chem. 1979, 8 3 , 2714. (16) Besley, L.; Bottomley, 0. A. J. Chem. Thermodyn. 1973, 5,397. (17) Douslin, D. R. J. Chem. Thermodyn. 1971, 3 , 187. (18) Stimson, H. F. J. Res. Natl. Bur. Stand. 1989. 73A, 493. (19) Osborne, N. S.; Stimson, H. F.; Flock, E. F.; Glnnlngs, D. C. J. Res. Natl. Bur. Stand. 1939. 23, 197. (20) Prytz, K. K. Dan. Vklensk. Selsk., Mat.-Fys. Medd. 1931, 2 , 1. (21) Johnson, D. P.; OulMner, L. A,; Jones, F. E. Proc. 8th Int. Conf. Prop. Steam Water, Glens, France 1974, Vol. I , 365. (22) Haar, L.; Gallagher, J.; Keil, 0. S. NBSINRC Steam Tables; Hemlsphere: New qork, 1984. (23) Wagner, W.; Polak, R. Proc. 8th Int. Conf. Prop. Water Steam, Glens, France 1874, Vol. 11, 787. (24) Wagner, W.; Ewers, J.; Penterman, W. J . Chem. Thermodyn. 1978, 8 . 1049. (25) Ambrose, D.; Lawrenson, I. J. J. Chem. T h e r m w n . 1972, 4 , 755. (26) Tanishta, I.; Watanabe, K.; Uematsu, M.; Eguchl, K. Proc. 8th Int. Conf. Prop. Water Steam, Glens, France 1974, Vol. I , 560. (27) TRC Thermodynamic Tables Nonhydrocarbons ; Thermodynamics Research Center, Texas ABM University; College Station, TX, 1985. (28) Prydz. R.; Goodwin, R. D. J. Chem. Thermodyn. 1972, 4 , 127. (29) Hoge, H. J. Res. Natl. Bur. Stand. 1950, 4 4 , 321. (30) Ambrose, D.; Sprake, C. H. S.; Townsend, R. J. Chem. Thermodyn. 1969, 7 , 499. (31) Ambrose, D. J. Chem. Thermodyn. 1981, 73, 1161. (32) Ambrose, D.; Broderick, B. E.; Townsend, R. J. Chem. SOC. 1987, 633. (33) Bender, P.; Furukawa. G. T.; Hyndman, J. R. Ind. Eng. Chem. 1952, 4 4 , 387. (34) Jackowski, A. W. J. Chem. Thermodyn. 1974, 6 , 49.

-

(35) Reid, R. C.; Prausnitz. J. M.; Sherwood, T. K. The Properties of Gases and Liquids, 3rd ed.;McGtaw-Hill: New York, 1977. (36) Lydersen, A. I. Estimation of Crltlcal Properties of Organlc Compounds; Unlverslty of Wisconsin: Madison, WI, 1965. Unlv. Wls.. Coil. Eng., Eng. Exp. Stn. Repr. 3. (37) Kreglewski, A. Bull. Acad. Pol. Sci., Ser. Sci. Chim. 1961, 9 , 163. (38) Kreglewski, A.; Zwolinskl, B. J. Rocz. Chem. 1968, 35, 1041, 1059. (39) Riedel, L. Chem. Ing. Tech. 1954, 2 6 , 83. (40) Kudchadker, A. P.; Alanl. G. H.; Zwolinskl. B. J. Chem. Rev. 1988, 6 8 , 659. (41) Ambrose, D.; Cox, J. D.; Townsend, R. Trans. faraday SOC. 1980, 56, 1. (42) Ambrose, D.; Patel, N. C. J. Chem. Thermodyn. 1984, 76, 459. (43) Ambrose, D. NPL Report Chem. 92, National Physical Laboratory, Teddlngton, Middlesex. UK, 1978. (44) Ambrose, D. NPL Report Chem. 98, National Physical Laboratory, Teddlngton, Middlesex, UK, 1979. (45) Kay, W. B. J. Am. Chem. SOC. 1948, 68, 1336. (46) McMicking, J. H.; Kay, W. 6. Proc., Am. Pet. Inst., Sect 3 1965, 45. (47) McGarry, J. Ind. Eng. Chem. Process Des. Dev. 1983, 22, 313. (48) Pitzer, K. S.J. Am. Chem. SOC. 1955, 7 7 , 3427. (49) Goodwin, R. D. J. Phys. Chem. Ref. Data 1985, 74, 849. (50) Goodwin, R. D. Cryogenics 1983, 2 3 , 403. (51) Ambrose. D.: SDrake. C. H. S.; Townsend, R. J. Chem. Thermodyn. 1972, 4 , 247. (52) Ambrose, D.; Sprake, C. H. S. J. Chem. Thermodyn. 1970, 2 , 631. (53) Ambrose, D.; Sprake, C. H. S.; Townsend, R. J. Chem. Thermodyn. 1875 . .-, 7 . , 185 . - -. (54) Ambrose, D.; Townsend, R. J. Chem. SOC. 1963, 3614. ( 5 5 ) Egan, C. J.; Kemp, J. D. J. Am. Chem. SOC. 1937, 59, 1265. (56) Douslln, D. R.; Harrison, R. H. J. Chem. Thermodyn. 1976, 8 , 301. (57) Kemp, J. D.; Giaque, W. F. J. Am. Chem. SOC. 1937, 59,79. (58) Willingham. C. 6.; Taylor, W. J.; Pignocco, J. M.; Rosslni, F. D. J . Res. Natl. Bw. Stand. 1945, 35,219. (59) Carruth, G. F.; Kobayashi, R. J. Chem. Eng. Data 1973, 18, 115. (60) Aston, J. G.; Messerly, G. H. J. Am. Chem. SOC. 1940, 62, 1917. (61) Kratzke, H.; Splllner, E.; Muller, S.J. Chem. Thermodyn. 1982, 74, 1175. (62) Loomis, A. G.;Waiters, J. E. J. Am. Chem. SOC. 1928. 4 8 , 2051. (63) Douslln, D. R.; Harrison, R. H. J. Chem. Thermodyn. 1973, 5, 491. (64) Kratzke, H. J. Chem. Thermodyn. 1980, 72, 305. (65) Kemp, J. D.; Egan, C. J. J. Am. Chem. SOC. 1938, 6 0 , 1521. (66) Hugill, J. A.; McGlashan, M. L. J. Chem. Thermodyn. 1978, IO, 95. (67) Robinson, D. B.; Senturk, N. H. J. Chem. Thermodyn. 1979, 7 1 , 461. (68) Ambrose, D. Trans. Faraday SOC. 1958, 52, 772. (69) Ambrose, D.; Townsend, R. Trans. faraday SOC. 1964, 6 0 , 1025. (70) Ambrose, D.; Sprake, C. H. S.;Townsend, R. J. Chem. Thermodyn. 1974, 6 , 693. (71) Ambrose, D.; Ellender, J. H.; Gundry, H. A.; Lee, D. A,; Townsend, R. J. Chem. Thermodyn. 1981, 13, 795. (72) Ambrose, D.; Ellender, J. H.; Sprake, C. H. S.;Townsend, R. J. Chem. Thermodyn. 1977, 9 ,735. (73) Thomas, H. P.; Harrison, R. H. J. Chem. Eng. Data 1982, 2 7 , 1. (74) Ambrose, D.; Hall, D. J.; Lee, D. A.; Lewis, G. B.; Mash, C. J. J. Chem. Thermodyn. 1978, 1 7 , 1089. (75) Vidaurri, F. C. J. Chep. Eng. Data 1975, 20. 349. (76) Ambrose, D.; Ellendei, J. H.;Sprake, C. H. S.J. Chem. Thermodyn. 1974, 6 , 909. (77) Osborn, A. G.; Douslin, D. R. J. Chem. Eng. Data 1988, 13, 534.

-

Received for review December 2, 1985. Revised May 30, 1986. Accepted June 19, 1986. This work Is financially supported by the Thermodynamics Research Center.